Purpose: The purpose of this study is to determine if illicit drug use or use of controlled substances are associated with an increased risk of gastrointestinal bleeding. Methods: We performed a retrospective analysis of 98 cases with corresponding matched controls presenting to the Birmingham Veterans Affairs Medical Center between July 2003 through April 2011. Using our electronic medical records system, data was collected on patients presenting with upper and lower gastrointestinal bleeding who had a urine drug screen performed upon presentation. Matched controls were selected from patients who were admitted for alternate reasons within 48 hours of each of the cases. These matched controls also had a urine drug screen obtained upon admission. The individual drug types that were recorded were marijuana, opioids, benzodiazepines, and cocaine. In addition, we collected data on BUN, creatinine, hematocrit, INR, current medication use, length of hospital stay, treatment type, and date of death in comparing both groups. The study has an 80% power to detect a change of 15% in UDS positivity between the groups with a p=0.05. Using SPSS, we analyzed the data using a chi-square 2-sided model to establish if UDS positivity confers an increased risk of gastrointestinal bleeding. Results: Medical records were reviewed for 98 cases and 98 controls that met study criteria. Nineteen of 98 (19.3%) cases of gastrointestinal bleeding were urine drug screen (UDS) positive for opioids, while 11 of 98 (11.2%) of our controls had a UDS that was positive for opioids (p=0.112). Twelve of 98 (12.2%) cases were UDS positive for cocaine, while 16/98 (16.3%) of the controls had a UDS that was positive for cocaine (p=0.414). Furthermore, 14/98 (14.2%) of the cases had a UDS that was positive for THC, whereas 9/98 (9.2%) of the controls had a UDS positive for THC (p=0.267). Ten of 98 (10.2%) cases were UDS positive for benzodiazepines, while 8 of 98 (8.2%) of the controls had a UDS positive for benzodiazepines (p=0.621). Overall, 37 of 98 cases (37.8%) had a positive UDS and 38 of 98 (38.8%) of control admissions had a positive UDS (p=0.883). Conclusion: There is no significant association between individual types of drug use or positive UDS and risk of gastrointestinal bleeding.Figure: [267]